HOW IT WORKS

AutoPilot Total Experience Control

Overview

In today’s media-saturated world, viewers have many choices for accessing content from multiple services/apps across their many devices. In this environment, over-the-top content providers struggle to keep subscribers engaged with their service. To reduce churn, there is a pressing need to differentiate from competing content providers and to capture the viewer’s loyalty.

Inspired by the proven success of Netflix, Junction TVs AutoPilot maximizes subscriber engagement and delivers total experience control. Moving above and beyond traditional media recommendations, AutoPilot automates the personalization of the entire screen experience – composed of recommended media, recommended meta-genres, personalized search, contextualized content, shelf programming, shelf placement and ordering, and more.

Leveraging cutting-edge cloud technologies, JunctionTV scales massively to analyze viewing data and content tags for each subscriber, auto-generating recommendations individually. AutoPilot employs multiple real-time adaptive control algorithms to map time-varying user tastes into a dynamic personalized experience.

Architecturally, AutoPilot is open and extensible via a micro-service architecture deployed in AWS. AutoPilot is available as a service via simple APIs that can be easily integrated with any application.

Auto Meta-Genre Creation

AutoPilot uses machine learning to create meta-genres based on the content tags and viewing history of individual viewers. JunctionTV’s meta-genres, which go beyond classical categories such as action, comedy or thrillers, allow for a more fine-grained representation of user tastes. For example, a viewer maybe interested only in “romantic comedies featuring Tom Hanks” and not just any comedy.

Personalized recommended shelf

AutoPilot is not limited to just one “Recommended for you” list. It can be used to create a total personalized experience on screen. It does so by building multiple recommendation shelves that reflect a deep understanding of the viewer’s taste. AutoPilot employs multiple shelf recommendation algorithms, such as “Recommended for you”, “Trending”, “Related to Movie xyz”, “Meta-genres: Movies casting Tom Hanks”, “Spotlights” etc.

Personalized search

Search results for individual users are automatically personalized, which includes user specific ranking of search results based on the user’s taste and viewing pattern.

Contextualization and filtering

AutoPilot is not limited to just one “Recommended for you” list. It can be used to create a total personalized experience on screen. It does so by building multiple recommendation shelves that reflect a deep understanding of the viewer’s taste. AutoPilot employs multiple shelf recommendation algorithms, such as “Recommended for you”, “Trending”, “Related to Movie xyz”, “Meta-genres: Movies casting Tom Hanks”, “Spotlights” etc.

Adaptive Shelf/Carousel Programming

Maximizing precious screen space is essential to an engaging experience. Viewers are frustrated with too much scrolling for content. By tracking user behavior, AutoPilot automates the entire process of shelf placement and selection. user behavior. The underlying adaptive shelf programming algorithm maximizes the shelf value in terms of viewing engagement.

Adaptive Content placement

When searching for content, a user’s attention span is limited. Content placement within a shelf is key to effective content discovery. AutoPilot drives automatic content placement by employing a ranking algorithm that is personalized for each user.